A number of treatments have been developed to treat articulatory deficits in acquired apraxia of speech (AOS). One such intervention is sound production treatment (SPT), which employs a response-contingent hierarchy made up of verbal modeling/repetition, graphic cueing, integral stimulation, and articulatory placement instruction. Interventions such as SPT are generally found to improve patients' speech accuracy but to varying degrees. In the current study, we will test the hypothesis that lesion site plays a role in the variability of SPT response. We will use state-of-the-art MRI imaging and post-processing techniques to better understand which patients are best able to benefit from AOS treatment, as well as the neural changes underlying treatment effects. Specifically, we will use voxel-based lesion symptom mapping to identify lesion sites most predictive of the degree of treatment response and advanced diffusion imaging techniques to map changes in the integrity of white matter tracts from pre- to post-treatment. To do this, we will recruit 30 individuals with AOS from the VA Northern California patient populations over the course of four years, using strict inclusion and exclusion criteria. Participants will be scanned on a 3T MRI scanner on the VA Martinez campus. Structural MRI and diffusion imaging sequences will be collected at three time points: before and after the baseline phase of treatment (to measure scan-rescan reliability prior to treatment) and again following treatment (to measure the neural changes associated with treatment). Treatment will be delivered by licensed speech-language pathologists familiar with this patient population and trained on the sound production treatment protocol. Patients will be trained on three different sounds (e.g., consonant, vowel sounds) that produce error rates >70% on pre-testing. Test stimuli will include both trained and untrained sounds to test for generalizability. To identify brin regions predictive of response to the SPT treatment, we will run voxel-based lesion symptom mapping analyses on normalized lesion reconstructions, with the independent variable being lesion status (lesioned or not) and the dependent variable, the change in sound production accuracy, with separate analyses for trained and untrained items. Based on our previous work and the relevant literature, we predict that the size of the SPT response (effect size) for both treated and untreated items will be affected by lesions involving the left superior pre-central gyrus of the insula. Next, diffusion imaging-derived tractography will be used to reconstruct eight major white matter pathways, and fractional anisotropy values will be calculated for specific regions of interest in each pathway. The degree of treatment response to both trained and untrained stimuli is predicted to correlate with fractional anisotropy values in the white matter surrounding the superior pre-central gyrus of the insula, the left arcuate, superior longitudinal, and medial subcallosal fasciculi, as these tracts have been implicated in studies of fluency and motor speech disorders. To measure neuroplasticity associated with the SPT protocol, we will calculate the difference in fractional anisotropy values from Scan 2 to 3 for each of the eight fiber tracts. Our predictions again focus on those tracts that have been directly associated with speech production in the literature, the arcuate, superior longitudinal, and medial subcallosal fasciculi, and fibers of the extreme capsule. In short, the current proposal is significant in thatit addresses a large gap in the literature with respect to prospectively identifying specific grey and white matter correlates that predict AOS treatment response, as well as the neural mechanisms underlying such treatment. The proposed project will improve our scientific understanding of the mechanisms underlying speech-language treatment and provide clinicians with critical information that they can use to improve patient care.